\begin{table}[H] \centering
\newcolumntype{R}{>{\raggedleft\arraybackslash}X}
\newcolumntype{L}{>{\raggedright\arraybackslash}X}
\newcolumntype{C}{>{\centering\arraybackslash}X}

\caption{Effects of electoral turnovers on governance and corruption}
\label{tab:baseline_het_corruption}
\begin{tabularx}{\linewidth}{@{}lCCCCCCC@{}}

\toprule
 & &  \multicolumn{2}{c}{\textbf{Election type}} & \multicolumn{2}{c}{\textbf{Checks \& balances}} & \multicolumn{2}{c}{\textbf{Leader power}} \tabularnewline & (1) & (2) & (3) & (4) & (5) & (6) & (7) \tabularnewline 
{}&{Baseline}&{Pres.}&{Parl.}&{High}&{Low}&{High}&{Low} \tabularnewline
\midrule \addlinespace[\belowrulesep]
Government accountability index&0.198**&0.205&0.180&0.043&0.333**&0.274*&-0.096 \tabularnewline
&(0.101)&(0.258)&(0.123)&(0.051)&(0.184)&(0.175)&(0.135) \tabularnewline
(Minus) Executive corruption index&0.276**&0.426&0.206&0.118&0.533**&0.265&0.186* \tabularnewline
&(0.133)&(0.309)&(0.138)&(0.094)&(0.259)&(0.289)&(0.127) \tabularnewline
(Minus) Public sector corruption index&0.163*&0.424*&0.062&-0.014&0.311**&0.251&-0.035 \tabularnewline
&(0.096)&(0.262)&(0.092)&(0.069)&(0.176)&(0.209)&(0.115) \tabularnewline
Control of corruption&0.442**&0.648&0.352&0.114&0.668**&1.015***&0.119 \tabularnewline
&(0.212)&(0.423)&(0.272)&(0.276)&(0.318)&(0.358)&(0.288) \tabularnewline
Absence of corruption index&0.240***&0.383&0.171**&0.061&0.399***&0.345**&-0.001 \tabularnewline
&(0.093)&(0.255)&(0.087)&(0.052)&(0.163)&(0.184)&(0.097) \tabularnewline
\bottomrule \addlinespace[\belowrulesep]

\end{tabularx}
\\ \parbox{\linewidth}{\footnotesize \caption*{\footnotesize \emph{Notes}: This table reports RD estimates corresponding to equation (\ref{eq:baseline_rdd}) for our measures of governance and corruption, expressed in standard deviation terms, in the main sample (column 1) and in different subsamples (columns 2 to 7). See Table \ref{tab:baseline_het_pres_parl_ballot} for the definition of the high and low checks and balances subsamples, and Appendix Table \ref{tab:baseline_het_global_checks_power} for the definition of high and low leader power subsamples. Using the method of \cite{clogg1995statistical}, we can reject the equality of the estimates for the absence of corruption index for presidential and parliamentary elections (p-val. =  0.431), but not for high and low checks and balances (p-val. =  0.048), and for high and low leader power (p-val. =  0.096). We obtain broadly consistent results when running a parametric regression in which we include the interaction between the treatment and the dimension of heterogeneity. $^{*} p<0.10,^{**} p<0.05,^{***} p<0.01$.}}
\end{table}
